Management system and method for highway dynamic charging

Abstract

The invention discloses a management system and method for highway dynamic charging, and relates to the technical field of intelligent traffic management. By setting different charging standards, drivers are stimulated to adjust travel, active adjustment of traffic flow is realized, users are fundamentally guided to select travel modes according to real-time road conditions and cost benefits, a traditional passive traffic diversion mode is changed, expressway network flow distribution is optimized, traffic jam is effectively relieved, and traffic safety is improved. The overall road network operation efficiency and service quality are improved; according to the invention, parallel prediction operation is carried out by using the flow prediction model to obtain the multi-objective optimization result of the highway dynamic toll collection scheme, including the optimal income target scheme, the optimal traffic flow target scheme, the road network balance target scheme and the minimum congestion time target scheme, so that multiple toll collection scheme selections are provided for an administrator, and the toll collection efficiency is improved. And different management requirements are met.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent traffic management, and particularly relates to a management system and method for dynamic toll collection on expressways. BACKGROUND

[0002] With the acceleration of urbanization and the continuous growth of car ownership, traffic congestion has become a common problem faced by major cities around the world. Traffic congestion not only leads to a significant increase in travel time, reducing people's quality of life, but also causes energy waste and environmental pollution, posing a serious threat to the sustainable development of cities. In this context, intelligent traffic management technology has emerged and gradually become a key means to solve traffic problems.

[0003] In order to alleviate traffic congestion on expressways, some areas have begun to implement dynamic toll collection strategies, i.e., adjusting toll standards according to different time periods, road sections, and traffic flow conditions, to guide drivers to avoid congestion periods and road sections. However, existing dynamic toll collection technologies have the following problems and deficiencies: Most current dynamic toll collection schemes are based on experience or simple statistical methods, lacking comprehensive consideration of factors such as traffic flow, driver behavior, road network structure, etc. This rough toll scheme formulation method is difficult to achieve precise control of traffic flow and cannot achieve the expected congestion relief effect.

[0004] In addition, existing dynamic toll collection technologies usually only focus on a single target, such as increasing toll revenue or alleviating traffic congestion, while ignoring other important targets such as road network balance, minimum congestion time, etc. In actual traffic management, these targets are often interrelated and interdependent, and simply pursuing a certain target may worsen other targets.

[0005] In view of the above problems, it is necessary to propose a management system and method for dynamic toll collection on expressways. SUMMARY

[0006] The purpose of the present application is to solve the problems in the background art and propose a management system and method for dynamic toll collection on expressways.

[0007] In order to achieve the above purpose, the present application adopts the following technical solutions: The present application provides, in a first aspect, a management system for dynamic toll collection on expressways, comprising a data acquisition module, a road network modeling module, a data analysis and traffic prediction module, a toll strategy matching and adjustment module, and a signal matching and execution module.

[0008] The data collection module collects the expressway dynamic data on the expressway in real time, including traffic flow data, fee collection data, road condition data, weather environment data, and path selection preference data of private cars, shared cars and public transport passengers facing differential charges.

[0009] The traffic flow data is collected, including daily average vehicle speed, daily average vehicle flow, daily average vehicle density, congestion length proportion and daily average accident rate of each road section. The fee collection data is collected, including basic fee and unit kilometer charge of each road section.

[0010] The road condition data is collected, including maintenance length proportion, average flatness and accumulated water road section length proportion of each road section.

[0011] The weather environment data is collected, including average temperature, average wind speed and daily average rainfall of each road section.

[0012] The path selection preference data of private cars, shared cars and public transport passengers facing differential charges is collected, including private car proportion, shared car proportion and public transport vehicle proportion of each road section.

[0013] The road network modeling module obtains the topological structure of each road section in the entire road network, and obtains the road length of each road section according to the topological structure. According to the topological structure of each road section and the expressway dynamic data, mathematical modeling is carried out, and the road length, road condition data and weather environment data of each road section are converted into a road network environment data matrix; the traffic flow data and the path selection preference data of private cars, shared cars and public transport passengers facing differential charges are converted into a road network road condition response data matrix; and the fee collection data is converted into a price control data matrix.

[0014] In the road network environment data matrix, the road network road condition response data matrix and the price control data matrix, each row represents a road section, and each column represents an expressway dynamic data. The generated road network environment data matrix, road network road condition response data matrix and price control data matrix are sent to the data analysis and flow prediction module.

[0015] The data analysis and flow prediction module extracts the correlation features and mathematical regression features between the road network environment data matrix, the road network road condition response data matrix and the price control data matrix through a machine learning algorithm based on a multilayer perceptron, and obtains a flow prediction model for predicting traffic flow, vehicle speed and vehicle type through expressway dynamic data.

[0016] As a preferred mode of the present application, a machine learning model based on a multilayer perceptron is established, including an input layer, a hidden layer and an output layer.

[0017] The input layer has a preset even number of nodes, and the first half of the nodes receive the road network environment data matrix, and the second half of the nodes receive the price control data matrix.

[0018] The hidden layer has a preset number of layers, and the input of the previous layer is the output of the next layer. Each node in the hidden layer processes the output result of the previous node through convolution operation, adds a bias term and an activation function, and increases the expression ability of the model through the activation function.

[0019] The output layer is used to predict the prediction result of the road network condition response data matrix from the current time to a preset number of days.

[0020] As a preferred mode of the present application, the multi-layer perception-based machine learning model is trained to determine the specific values of the weight matrix and the bias term in the output layer and each hidden layer, and the specific process is as follows: The loss function is defined, the error between the predicted result of the road network condition response data matrix and the true result is described by calculating the average value of the sum of the squares of the difference between the element values in the output layer and the corresponding true result, and the loss value is obtained.

[0021] A preset number of training samples, i.e. the road condition response data matrix at each time according to the real data, and the corresponding prediction result at each time are obtained. For each training sample, forward propagation calculation is performed in turn, and the gradient of the loss function with respect to the weight matrix of the output layer and each hidden layer is calculated by the back propagation algorithm based on the loss function: In the back propagation calculation process, the weight matrix and the bias term are updated by the stochastic gradient descent method SGD, and in the update process, the first, second, third and fourth learning rates are used as the influence coefficients of the loss function with respect to the hidden layer weight matrix, the hidden layer weight matrix bias term, the output layer weight matrix and the output layer bias term, to control the update of the weight matrix and the bias term.

[0022] The weight matrix and the bias term are updated and iterated by the stochastic gradient descent method SGD until the termination condition is reached: the specific value of the loss function is less than the preset threshold. At this time, the final updated specific values of the hidden layer weight matrix, the hidden layer weight matrix bias term, the output layer weight matrix and the output layer bias term are recorded, and are brought back to the corresponding number of layers and node positions in the hidden layer and the output layer to obtain the final traffic prediction model.

[0023] The toll policy matching and adjusting module carries out parallel prediction operation through the traffic prediction model, obtains multi-objective optimization results representing the dynamic toll scheme of the expressway, including an optimal revenue target scheme, an optimal traffic flow target scheme, a road network balance target scheme and a minimum congestion time target scheme, and recommends the dynamic toll scheme of the expressway according to the multi-objective optimization results.

[0024] As a preferred mode of the present application, the parallel prediction operation is carried out based on the traffic prediction model, the traffic prediction model is copied several times, and the copied traffic prediction models are sequentially numbered.

[0025] The upper and lower limits of the basic fee and the unit kilometer toll of each road section are defined, a random array uniquely corresponding to the copied traffic prediction model is randomly generated between the upper and lower limits of the basic fee and the unit kilometer toll of each road section, and a price control data matrix representing a randomly generated simulation value of the toll is obtained.

[0026] The randomly generated price control data matrix and the collected road network environment data matrix historical data are combined to form a data set for inputting the input layer of the traffic prediction model, and all the data sets are respectively input into the traffic prediction models uniquely corresponding thereto, and the prediction results of the traffic prediction models are obtained.

[0027] Elements are extracted from the prediction results of the traffic prediction models, and the average value of the daily average speed, the average value of the traffic flow, the average value of the daily average driving density, the average value of the congestion length proportion, the average value of the daily average accident rate, the average value of the private car proportion, the average value of the shared car proportion and the average value of the public transport vehicle proportion of each road section in the road network road condition response data matrix are calculated.

[0028] As a preferred mode of the present application, the multi-objective optimization results, including the optimal revenue target scheme, the optimal traffic flow target scheme, the road network balance target scheme and the minimum congestion time target scheme, are found in the final results of each prediction through the target function, and the specific process is as follows: The optimal revenue target scheme target function is defined: the product of the average value of the daily average speed of each road section and the unit kilometer toll is added to the basic fee of each road section to obtain result one, result one is multiplied by the average value of the traffic flow to obtain result two, and the traffic prediction model number satisfying the condition that the sum of result two of each road section is maximum is solved.

[0029] The optimal traffic flow target scheme target function is defined according to the average value of the daily average speed, the average value of the congestion length proportion and the average value of the daily average accident rate output by each traffic prediction model; The traffic prediction model number satisfying the condition that the optimal traffic flow target scheme target function result is maximum is solved.

[0030] Define the road network equilibrium target scheme target function: Solve the square root of the average daily vehicle speed of each road section about the road section, get result four; solve the square root of the difference between the average traffic flow and the length of the road section about the road section, get result five; solve the square root of the average of each daily average vehicle density about the road section, get result six; solve the flow prediction model number that meets the condition that the sum of result four, result five and result six of each road section is the smallest; Define the minimum congestion time target scheme target function: Calculate the sum of the average of the congestion length proportion of each road section and the average of the daily average accident rate, get result seven; solve the flow prediction model number that meets the condition that the sum of result seven about the road section is the smallest.

[0031] Get the prediction result numbers that meet the best income target scheme target function, the best traffic flow target scheme target function, the road network equilibrium target scheme target function and the minimum congestion time target scheme target function respectively, denoted as the first number, the second number, the third number and the fourth number.

[0032] The signal matching and executing module sends the multi-objective optimization result representing the dynamic toll scheme of the expressway to the administrator, recommends the scheme, and obtains the expressway dynamic toll scheme selected by the administrator. According to the expressway dynamic toll scheme selected by the administrator, dynamic toll adjustment is implemented. Verify the consistency of the price control data matrix corresponding to each multi-objective optimization result, and according to the consistency verification result, feedback correction is made to the operating parameters in the flow prediction model, and the prediction result of the flow prediction model is optimized.

[0033] The best income target scheme, the best traffic flow target scheme, the road network equilibrium target scheme and the minimum congestion time target scheme are sent to the administrator, and the corresponding prediction results, i.e. the road network condition response data matrix output by the flow prediction model numbered as the first number, the second number, the third number and the fourth number are sent to the administrator, and dynamic toll adjustment is implemented according to the expressway dynamic toll scheme selected by the administrator.

[0034] As a preferred mode of the present application, the consistency of each multi-objective optimization result corresponding to the price control data matrix is verified, the input price control data matrix in the traffic prediction model numbered as the first number, the second number, the third number and the fourth number is obtained, and the consistency coefficient between each price control data matrix is obtained through the sum of the absolute value of the difference of each same position element between each price control data matrix. When there is a consistency coefficient less than a preset threshold, it is considered that there is a similar price control data matrix between the best benefit target scheme, the best traffic flow target scheme, the road network equilibrium target scheme and the minimum congestion time target scheme, it is determined that the traffic prediction model training is insufficient, and the multi-objective optimization result is insufficient in discrimination.

[0035] If it is identified that the multi-objective optimization result is insufficient in discrimination, the data analysis and traffic prediction module is commanded to retrain the traffic prediction model training, and the training parameters: the first, second, third and fourth learning rates are reduced to half of the initial value.

[0036] The present application provides a management method of highway dynamic charging in a second aspect, comprising the following steps: Step one, data acquisition and formatting processing; Real-time acquisition of highway dynamic data on the highway, including traffic flow data, fee collection data, road condition data, weather environment data and path selection preference data of private car, shared car and public transportation passengers facing differential charging. Obtain the topological structure of each road section in the entire road network, and obtain the length of each road section according to the topological structure. According to the topological structure of each road section and the highway dynamic data, mathematical modeling is carried out, and the length of each road section, road condition data and weather environment data are converted into road network environment data matrix; the traffic flow data and the path selection preference data of private car, shared car and public transportation passengers facing differential charging are converted into road network traffic response data matrix; the fee collection data is converted into price control data matrix.

[0037] Step two, machine learning modeling and training; Feature extraction and regression fitting are carried out through a machine learning algorithm based on multilayer perception, the correlation features and mathematical regression features between the road network environment data matrix, the road network traffic response data matrix and the price control data matrix are extracted, and the traffic prediction model for predicting traffic flow, vehicle speed and vehicle type through highway dynamic data is obtained. The future traffic flow trend and the selection trend of vehicle type are predicted through the traffic prediction model, and the prediction result data is obtained, which provides basic data for traffic flow optimization and charging mechanism change decision.

[0038] Step three, multi-objective prediction based on machine learning; The optimization operation is carried out through the traffic prediction model, the multi-objective optimization result representing the expressway dynamic charging scheme is obtained through parallel prediction operation and the target function, including the best income target scheme, the best traffic flow target scheme, the road network balance target scheme and the minimum congestion time target scheme, and the expressway dynamic charging scheme recommendation is carried out according to the multi-objective optimization result.

[0039] Step four, prediction result verification and execution; The multi-objective optimization result representing the expressway dynamic charging scheme is sent to the administrator, the scheme recommendation is carried out, and the expressway dynamic charging scheme selected by the administrator is obtained. The dynamic charging adjustment is implemented according to the expressway dynamic charging scheme selected by the administrator.

[0040] Step five, machine learning parameter feedback adjustment; In the implementation process of the dynamic charging adjustment, the consistency of the actual generated traffic flow data and the path selection preference data of private cars, shared cars and public transport passengers facing different charges with the prediction result of the expressway dynamic charging scheme selected by the administrator is verified. According to the consistency verification result, the running parameters in the traffic prediction model are feedback corrected, and the prediction result of the traffic prediction model is optimized.

[0041] The beneficial effects of the present application are: 1. The present application stimulates drivers to adjust travel by setting different charging standards, actively adjusts traffic flow, fundamentally guides users to choose travel mode according to real-time road conditions and cost-effectiveness, changes the traditional "passive" traffic distribution mode, optimizes the traffic distribution of the expressway network, effectively relieves traffic congestion, and improves the overall road network operation efficiency and service quality; 2. The present application uses the traffic prediction model to carry out parallel prediction operation, obtains the multi-objective optimization result of the expressway dynamic charging scheme, including the best income target scheme, the best traffic flow target scheme, the road network balance target scheme and the minimum congestion time target scheme, provides the administrator with multiple charging scheme selection, and meets different management needs; 3. In the implementation process of the dynamic charging adjustment, the consistency of the actual data and the prediction result is verified, the running parameters of the traffic prediction model are feedback corrected according to the verification result, when the multi-objective optimization result is found to be insufficient, the model is retrained and the training parameters are adjusted, so as to optimize the prediction result of the traffic prediction model. BRIEF DESCRIPTION OF DRAWINGS

[0042] Figure 1 The system architecture schematic diagram of the expressway dynamic charging management system provided in the embodiment of the present application; Figure 2 The machine learning model topology structure diagram based on the multi-layer perception machine provided in the embodiment of the present application; Figure 3 A method flow chart of a management method of a highway dynamic toll is provided in the embodiment of the present application. DETAILED DESCRIPTION

[0043] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all the other embodiments obtained by those skilled in the art without creative work are within the protection scope of the present application.

[0044] In addition, if the present application embodiments involve descriptions such as "first", "second", etc., the descriptions of "first", "second", etc. are only for the purpose of description, and cannot be understood as indicating or implying the relative importance of the indicated technical features or implicitly indicating the number of the indicated technical features. Therefore, the features limited by "first", "second" can be explicitly or implicitly included at least one of the features. In addition, the meaning of "and / or" appearing throughout the text includes three parallel solutions. For example, "A and / or B" includes A solution or B solution, or A and B solutions that meet at the same time. In addition, the technical solutions of each embodiment can be combined with each other, but it must be based on the fact that a person skilled in the art can realize it. When the combination of technical solutions appears contradictory or unachievable, it should be considered that the combination of technical solutions does not exist, and is not within the protection scope claimed by the present application.

[0045] Please refer to Figure 1 As shown in the figure, a highway dynamic toll management system includes a data acquisition module, a road network modeling module, a data analysis and traffic prediction module, a charging strategy matching and adjustment module, and a signal matching and execution module.

[0046] The data acquisition module is responsible for accessing sensors, API interfaces, cameras, and GPS devices, and collecting real-time highway dynamic data on the highway, including traffic flow data, fee collection data, road condition data, weather environment data, and collecting private car, shared car, and public transportation passenger path selection preference data for differential charging. The collected highway dynamic data is sent to the data analysis and traffic prediction module.

[0047] Traffic flow data is collected, including the daily average speed of each road segment , the daily average traffic flow , the daily average traffic density , the congestion length ratio , and the daily average accident rate ; Collecting the toll data, including the basic toll of each road segment i and the toll per kilometer .

[0048] Collecting the road condition data, including the maintenance length ratio of each road segment i , the average flatness and the length ratio of waterlogged road segment .

[0049] Collecting the weather environment data, including the average temperature , the average wind speed and the daily average rainfall of each road segment .

[0050] Collecting the path selection preference data of private cars, shared cars and public transportation passengers facing different charges, including the private car ratio , the shared car ratio and the public transportation vehicle ratio of each road segment .

[0051] wherein, is the road segment number, ; is the total number of road segments in the road network; wherein is the data collection time, in days.

[0052] The road network modeling module obtains the topological structure of each road segment in the entire road network, and obtains the length of each road segment according to the topological structure. According to the topological structure of each road segment and the dynamic data of the expressway, mathematical modeling is carried out to convert the length of each road segment, the road condition data and the weather environment data into a road network environment data matrix; the traffic flow data and the path selection preference data of private cars, shared cars and public transportation passengers facing different charges are converted into a road network condition response data matrix; and the toll data is converted into a price control data matrix.

[0053] Obtain the topological structure of each road segment in the entire road network, generate a two-dimensional model of each road segment in the road network, and obtain the length of each road segment in the two-dimensional model.

[0054] Further, according to the topological structure of each road segment and the dynamic data of the expressway, mathematical modeling is carried out, and the specific process is as follows: Generating a road network environment data matrix: , Generating a road network condition response data matrix: ,​ The price control data matrix is generated: , In the road network environment data matrix, the road network traffic response data matrix and the price control data matrix, each row represents a road segment ; each column represents a highway dynamic data.

[0055] The generated road network environment data matrix, road network traffic response data matrix and price control data matrix are sent to the data analysis and traffic prediction module.

[0056] The data analysis and traffic prediction module extracts the correlation features and mathematical regression features between the road network environment data matrix, the road network traffic response data matrix and the price control data matrix through a multi-layer perceptron-based machine learning algorithm, and obtains a traffic prediction model for predicting traffic flow, vehicle speed and vehicle type through highway dynamic data.

[0057] Referring to FIG. 8, a multi-layer perceptron-based machine learning model is established, including an input layer, a hidden layer and an output layer. Figure 2

[0058] The input layer has nodes, respectively receiving the data sets of the road network environment data matrix and the price control data matrix collected from the previous days: and , wherein is a preset model backtracking constant, representing the number of days of historical data window that needs to be backtracked for predicting the road network traffic response data matrix.

[0059] The hidden layer has a preset number of layers , and each layer has a preset number of nodes , wherein The input of the previous layer in the hidden layer is the output of the next layer. Each node in the hidden layer processes the output result of the previous node through convolution operation, adds a bias term and an activation function, and increases the expression ability of the model through the activation function.

[0060] The operation formula of the hidden layer is:

[0061] wherein is the layer index of the hidden layer, ; is the output value of the th node in the th hidden layer, wherein is an activation function.​ is the number of nodes in the first hidden layer, ; is the number of nodes in the first hidden layer, the total number of nodes in the hidden layer; wherein is the output of the i-th node in the j-th hidden layer, and are the weight matrix and bias term of the i-th node in the j-th hidden layer, respectively.

[0062] wherein the output layer has a preset number of nodes , respectively, for predicting the prediction results of the road network traffic response data matrix from the current time to days later:

[0063] The output formula of the output layer is:

[0064] wherein is the node index of the output layer, respectively, corresponding to the prediction results ; wherein is the output of the i-th node in the j-th hidden layer, , wherein and are the weight matrix and bias term of the i-th node in the output layer, respectively. Further, the multi-layer perceptron-based machine learning model is trained to determine the specific values of the weight matrix and bias term in the output layer and each hidden layer, and the specific process is as follows:

[0065] Define a loss function to describe the error between the prediction results and the true results of the road network traffic response data matrix: wherein

[0066] is the column number of the road network traffic response data matrix, . wherein

[0067] is the element value of the i-th row and the j-th column in the road network traffic response data matrix output by the output layer on the i-th day, wherein is the true result corresponding to the element value . ​​​​​​​​​​

[0068] A preset number of training samples, i.e. the road condition response data matrix at each time point obtained according to the real data, and the prediction results corresponding thereto at each time point are collected. For each training sample, forward propagation calculation is sequentially performed, and the gradients of the loss function with respect to the weight matrix and bias term of the output layer and each hidden layer are calculated through the back propagation algorithm: The gradient calculation formula of the weight matrix and bias term of the hidden layer is:

[0069] The gradient calculation formula of the weight matrix and bias term of the output layer is:

[0070] The weight matrix and bias term are updated by the stochastic gradient descent method (SGD), and the learning rate: the first learning rate , the second learning rate , the third learning rate and the fourth learning rate are used to control the update of the weight matrix and bias term, and the formula is: The gradient update formula of the weight matrix and bias term of the hidden layer is: , The gradient update formula of the weight matrix and bias term of the output layer is: , wherein , , and are the specific values of the hidden layer weight matrix, the hidden layer weight matrix bias term, the output layer weight matrix and the output layer bias term after update; wherein , , and are the original values of the hidden layer weight matrix, the hidden layer weight matrix bias term, the output layer weight matrix and the output layer bias term before update.

[0071] The weight matrix and bias term are updated by the stochastic gradient descent method (SGD) for iterative operation until the termination condition is reached: the specific value of the loss function is less than the preset threshold . At this time, the final updated specific values of the hidden layer weight matrix, the hidden layer weight matrix bias term, the output layer weight matrix and the output layer bias term are recorded, and are brought back to the corresponding layer and node position in the hidden layer and the output layer to obtain the final traffic prediction model.

[0072] ​​It's important to note that in modern traffic management, road pricing, as an effective means of managing traffic demand, can improve economic efficiency while optimizing road usage. The basic principle is to proactively regulate traffic flow by setting different toll rates to incentivize drivers to make adjustments during their trips. This approach fundamentally guides users to make travel choices based on real-time road conditions and cost-effectiveness, transforming the traditional "passive" traffic diversion model into a more flexible and efficient "active" regulation. This approach not only optimizes traffic flow distribution within the expressway network, but also effectively alleviates traffic congestion and improves the overall operational efficiency and service quality of the road network.

[0073] It's important to note that the traffic flow prediction model aims to use a multi-layer perceptron algorithm to learn patterns between road toll conditions and traffic flow from historical data and use these patterns to predict future traffic flow. The input layer receives road network conditions and price control data from the previous k1 days. This data is processed through convolution operations and activation functions in multiple hidden layers, and the output layer predicts road network flow for the next few days.

[0074] The charging strategy matching and adjustment module carries out parallel prediction operations through the traffic prediction model to obtain multi-objective optimization results representing the dynamic charging scheme of the highway, including the optimal revenue target scheme, the optimal traffic flow target scheme, the road network balance target scheme and the minimum congestion time target scheme, and recommends the dynamic charging scheme of the highway based on the multi-objective optimization results.

[0075] Based on the traffic prediction model, parallel prediction calculation is performed and the traffic prediction model is copied and numbered them ; Limit each road section The upper and lower limits of the basic fee and the per-kilometer fee are set for each section of the road. The base fare and the upper and lower limits of the unit kilometer are randomly generated Arrays: ,

[0076] Among them = , ,..., Respectively The randomly generated simulation value of the basic cost of each road section i in the array; = , ,..., Respectively In the array, each road segment randomly generated simulation values of the unit kilometer toll fees; The total of the generated randomly generated price control data matrices .

[0077] The randomly generated price control data matrices and the road network environment data matrices collected on the previous days are respectively composed into sets of data sets for input layers of the groups of traffic prediction models, including: The first group: and ; The second group: and ; The third group: and ;

[0078] The Pth group: and ; The price control data matrices , ,..., contained in the , ,..., , respectively, all have a quantity of .

[0079] The first group to the group of data sets are respectively put into the copied traffic prediction models according to their respective number symbols, to simulate the execution of the combination of basic fees and unit kilometer toll fees in consecutive days: The road network environment data: under the condition of producing road network traffic response data, to obtain the prediction results of each traffic prediction model .

[0080] From the prediction results of each traffic prediction model , the elements are extracted, and the average value of the average vehicle speed , the average value of the average vehicle flow in the total of days of road network traffic response data matrix about the average value of the average vehicle speed , the average daily traffic density , the average value of the congestion length ratio , the average daily accident rate , the average proportion of private cars , the average proportion of shared cars and the average proportion of public transport vehicles .

[0081] Furthermore, through the objective function in each prediction The final result Find the multi-objective optimization results, including the optimal revenue target solution, the optimal traffic flow target solution, the road network balance target solution and the minimum congestion time target solution. The specific process is as follows: Define the objective function of the optimal profit target plan: add the average value of the daily average vehicle speed of each section and the fee per kilometer to the basic fee of each section to obtain result one, multiply result one by the average traffic volume to obtain result two, and solve the condition: the traffic prediction model number when the sum of result two of each section is maximized.

[0082] The average value of the daily average vehicle speed output by each traffic prediction model p , the average value of the congestion length ratio and the average of the daily accident rate Define the objective function of the optimal traffic flow target solution:

[0083] Solve the conditions: Let the optimal traffic flow target solution objective function result: The flow prediction model number p when the result is the largest; Define the objective function of the road network balance target solution: Solve the sum of the square roots of the daily average vehicle speed for each road section to obtain result 4; solve the sum of the square roots of the difference between the average value of the traffic volume and the length of the road section to obtain result 5; solve the sum of the square roots of the average value of the daily average traffic density for each road section to obtain result 6; solve the traffic prediction model number that satisfies the condition: minimize the sum of result 4, result 5 and result 6 for each road section; Define the objective function of the minimum congestion time target solution: Calculate the sum of the average value of the congestion length ratio of each road section and the average value of the daily average accident rate to obtain result seven; solve the flow prediction model number that satisfies the condition: minimize the sum of result seven for the road section.

[0084] Obtain the predicted result numbers satisfying the optimal revenue target scheme objective function, the optimal traffic flow target scheme objective function, the road network equilibrium target scheme objective function and the minimum congestion time target scheme objective function respectively, denoted as the first number , the second number , the third number and the fourth number .

[0085] The signal matching and execution module sends the multi-objective optimization results representing the dynamic toll scheme of the expressway to the administrator, recommends the scheme, and obtains the expressway dynamic toll scheme selected by the administrator. According to the expressway dynamic toll scheme selected by the administrator, the dynamic toll adjustment is implemented. The consistency of the price control data matrix corresponding to each multi-objective optimization result is verified, and the running parameters in the traffic flow prediction model are feedback corrected according to the consistency verification result, so as to optimize the prediction result of the traffic flow prediction model.

[0086] The optimal revenue target scheme, the optimal traffic flow target scheme, the road network equilibrium target scheme and the minimum congestion time target scheme are sent to the administrator, and the corresponding prediction results, i.e. the road network traffic response data matrix output by the traffic flow prediction model numbered , , and are sent to the administrator, and the dynamic toll adjustment is implemented according to the expressway dynamic toll scheme selected by the administrator.

[0087] Further, the consistency of the price control data matrix corresponding to each multi-objective optimization result is verified, and the input price control data matrix in the traffic flow prediction model numbered , , and is obtained. , , and .

[0088] The consistency coefficient between each price control data matrix is obtained by calculating the sum of the absolute values of the differences of , , and each element, including: and consistency coefficient ; and consistency coefficient ; and The consistency coefficient ; and The consistency coefficient ; and The consistency coefficient ; and The consistency coefficient ; When the consistency coefficient is less than the preset threshold, it is considered that there are similar price control data matrices between the best profit target plan, the best traffic flow target plan, the road network balance target plan and the minimum congestion time target plan, and it is determined that the traffic prediction model is not sufficiently trained and the output multi-objective optimization results have insufficient discrimination.

[0089] If it is found that the multi-objective optimization result has insufficient discrimination, the data analysis and traffic prediction module is instructed to retrain the traffic prediction model and adjust the training parameters: the first learning rate , the second learning rate , the third learning rate and the fourth learning rate Reduced to half of the initial value.

[0090] See also Figure 3 As shown, a method for managing dynamic toll collection on expressways includes the following steps: Step 1: Data collection and formatting; Real-time data collection on expressways includes traffic flow, toll collection, road condition, weather, and environmental data, as well as data on the route preferences of private car, ride-sharing, and public transit passengers in response to differential tolls. The topology of each road segment in the entire road network is obtained, and the distance of each segment is determined based on this topology. Mathematical modeling is performed based on the topology of each segment and the expressway dynamics data, converting the distance, road condition, and weather data of each segment into a road network environment data matrix. Traffic flow data and the route preferences of private car, ride-sharing, and public transit passengers in response to differential tolls are converted into a road network road condition response data matrix. Toll collection data is converted into a price control data matrix.

[0091] Step 2: Machine learning modeling and training; Through feature extraction and regression fitting based on a multi-layer perceptron machine learning algorithm, the correlation features and mathematical regression features between the road network environment data matrix, the road network traffic response data matrix, and the price control data matrix are extracted, and a traffic flow prediction model for predicting traffic flow, vehicle speed, and vehicle type through the highway dynamic data is obtained. Through the traffic flow prediction model, the future traffic flow trend and the selection trend of vehicle types are predicted, and the prediction result data is obtained, which provides basic data for traffic flow optimization and toll mechanism change decisions.

[0092] Step three, multi-objective prediction based on machine learning; Through the traffic flow prediction model, the traffic flow prediction model is optimized, and the multi-objective optimization results representing the highway dynamic toll scheme are obtained through parallel prediction operation and objective function, including the best income target scheme, the best traffic flow target scheme, the road network balance target scheme, and the minimum congestion time target scheme, and the highway dynamic toll scheme is recommended according to the multi-objective optimization results.

[0093] Step four, prediction result verification and execution; The multi-objective optimization results representing the highway dynamic toll scheme are sent to the administrator for scheme recommendation, and the administrator selects the highway dynamic toll scheme. The dynamic toll adjustment is implemented according to the administrator's selection of the highway dynamic toll scheme.

[0094] Step five, machine learning parameter feedback adjustment; In the implementation process of dynamic toll adjustment, the consistency of the actual generated traffic flow data and the path selection preference data of private cars, shared cars, and public transportation passengers facing different charges with the prediction results of the administrator's selected highway dynamic toll scheme is verified. According to the consistency verification result, the operating parameters in the traffic flow prediction model are feedback corrected, and the prediction result of the traffic flow prediction model is optimized.

Claims

1. A dynamic toll collection management system for expressways, characterized in that: include: Data acquisition module, real-time collection of highway dynamic data on the highway; The road network modeling module obtains the topological structure of each road section in the entire road network and obtains the distance of each road section based on the topological structure; mathematical modeling is performed based on the topological structure of each road section and the dynamic data of the highway to obtain the road network environment data matrix, the road network traffic response data matrix and the price control data matrix; The data analysis and traffic flow prediction module uses a machine learning algorithm based on a multi-layer perceptron to extract the correlation characteristics and mathematical regression characteristics between the road network environment data matrix, the road network traffic response data matrix, and the price control data matrix to obtain a traffic flow prediction model; The charging strategy matching and adjustment module carries out parallel prediction operations through the traffic prediction model to obtain multi-objective optimization results representing the dynamic charging scheme of the highway.

2. A dynamic toll collection management system for expressways according to claim 1, characterized in that: Also includes signal matching and execution modules: The signal matching and execution module sends the generated multi-objective optimization results to the administrator, recommends solutions, and obtains the dynamic highway charging solution selected by the administrator; implements dynamic charging adjustments based on the dynamic highway charging solution selected by the administrator; verifies the consistency of the price control data matrix corresponding to each multi-objective optimization result, and provides feedback and corrections to the operating parameters in the traffic prediction model based on the consistency verification results.

3. A dynamic toll collection management system for expressways according to claim 1, characterized in that: The collected highway dynamic data include: Traffic flow data, including daily average vehicle speed, daily average traffic volume, daily average traffic density, congestion length ratio, and daily average accident rate for each road section; Fee collection data, including basic fees for each road section and fees per kilometer; Road condition data, including the proportion of each road section's maintenance length, average roughness, and the proportion of flooded sections; Weather data, including average temperature, average wind speed, and average daily rainfall for each road section; Data on route choice preferences of private car, shared car and public transport passengers facing differential tolls, including the proportion of private cars, shared cars and public transport vehicles on each road section.

4. A dynamic toll collection management system for expressways according to claim 1, characterized in that: In the road network environment data matrix, road network traffic condition response data matrix and price control data matrix, each row represents a road section; each column represents a highway dynamic data; the generated road network environment data matrix, road network traffic condition response data matrix and price control data matrix are sent to the data analysis and traffic prediction module.

5. A dynamic toll collection management system for expressways according to claim 1, characterized in that: The specific process of feature extraction and regression fitting through the machine learning algorithm based on the multi-layer perceptron is as follows: Build a machine learning model based on a multi-layer perceptron, including input layer, hidden layer, and output layer; Among them, the input layer is used to receive the historical data of the road network environment data matrix and the price control data matrix; The hidden layer has a preset number of layers, and the input of the previous layer is the output of the next layer. Each node in the hidden layer processes the output of the previous node through convolution operations, adding bias terms and activation functions, and the activation function increases the expressive power of the model. The output layer is used to predict the road network condition response data matrix from the current moment to the preset time; Train the machine learning model based on the multi-layer perceptron to determine the specific values ​​of the weight matrix and bias terms in the output layer and each hidden layer.

6. A dynamic toll collection management system for expressways according to claim 5, characterized in that: The specific process of training a machine learning model based on a multi-layer perceptron is as follows: Define the loss function, which describes the error between the predicted result and the actual result of the road network road condition response data matrix by calculating the average value of the sum of the squares of the differences between the element values ​​in the road condition response data matrix output by the output layer and the corresponding actual results, and obtain the loss value; Perform forward propagation calculation based on the loss value, and calculate the gradient of the loss function with respect to the weight matrix of the output layer and each hidden layer through the backpropagation algorithm based on the loss function, and update the weight matrix and bias term; During the back-propagation calculation process, the weight matrix and bias term are updated by the stochastic gradient descent method SGD, and the update speed is controlled by the first, second, third and fourth learning rates during the update process; The weight matrix and bias term are updated iteratively until the termination condition is reached: the specific value of the loss function is less than the preset threshold.

7. A dynamic toll collection management system for expressways according to claim 1, characterized in that: The specific process of launching parallel prediction operations through the traffic prediction model is as follows: Duplicate the traffic prediction model several times and number them sequentially; Limiting the upper and lower limits of the basic fee and the per-kilometer fee for each road section, randomly generating a random array uniquely corresponding to the copied traffic prediction model between the basic fee and the per-kilometer upper and lower limits for each road section, and obtaining a price control data matrix representing the randomly generated simulated values ​​of the fee; The randomly generated price control data matrix and the collected historical data of the road network environment data matrix are combined into a data set for the input layer of the traffic prediction model, and all data sets are input into the traffic prediction model corresponding to them respectively to obtain the prediction results of each traffic prediction model; Extract elements from the prediction results of each traffic prediction model and calculate the average daily average vehicle speed, average traffic volume, average daily average traffic density, average congestion length ratio, average daily average accident rate, average proportion of private cars, average proportion of shared cars, and average proportion of public transportation vehicles for each road section in the road network traffic response data matrix. Through the objective function, multi-objective optimization results are found in the final results of each prediction, including the best profit target solution, the best traffic flow target solution, the road network balance target solution and the minimum congestion time target solution.

8. A dynamic toll collection management system for expressways according to claim 7, characterized in that: The specific process of finding the multi-objective optimization result in the final results of each prediction through the objective function is as follows: Define the objective function of the optimal revenue target solution: add the product of the average daily vehicle speed of each road section and the fee per kilometer to the basic fee of each road section to obtain result 1, multiply result 1 by the average traffic volume to obtain result 2, and solve the traffic prediction model number that satisfies the condition: the sum of result 2 of each road section is maximized; Define the objective function of the optimal traffic flow target solution: The objective function of the optimal traffic flow target solution is defined based on the average daily vehicle speed, the average congestion length ratio, and the average daily accident rate output by each traffic flow prediction model; Solve the following conditions: the traffic flow prediction model number when the objective function result of the optimal traffic flow target solution is maximized; Define the objective function of the road network balance target solution: Solve the square root of the average daily vehicle speed for each road section and the sum of the results, and get result 4; solve the square root of the difference between the average value of the traffic volume and the length of the road section for each road section and the sum of the results, and get result 5; solve the square root of the average value of the average daily traffic density for each road section and the sum of the results, and get result 6; Find the traffic prediction model number that satisfies the condition: the sum of the results 4, 5 and 6 of each road section is minimized; Define the objective function of the minimum congestion time target solution: Calculate the sum of the average value of the congestion length ratio of each road section and the average value of the daily average accident rate to obtain result seven; solve the traffic prediction model number that satisfies the condition: the sum of result seven for each road section is minimized; Obtain the prediction result numbers that respectively meet the objective function of the optimal profit target scheme, the objective function of the optimal traffic flow target scheme, the objective function of the road network balance target scheme, and the objective function of the minimum congestion time target scheme, and record them as the first number, the second number, the third number, and the fourth number.

9. A dynamic toll collection management system for expressways according to claim 2, characterized in that: The specific process of optimizing the prediction results of the traffic prediction model is as follows: The generated optimal revenue target plan, optimal traffic flow target plan, road network balance target plan, and minimum congestion time target plan are sent to the administrator, and the corresponding prediction results, i.e., the road network traffic condition response data matrix output by the traffic prediction models numbered first, second, third, and fourth, are sent to the administrator; Obtain price control data matrices numbered as the first number, the second number, the third number, and the fourth number, and calculate the consistency coefficient between each price control data matrix; when there is a consistency coefficient less than a preset threshold, it is determined that the traffic prediction model is not sufficiently trained, and the output multi-objective optimization result has insufficient discrimination, and the command data analysis and traffic prediction module is commanded to retrain the traffic prediction model, and the training parameters: the first, second, third, and fourth learning rates are reduced to half of the initial value.

10. A method for managing dynamic toll collection on expressways, characterized in that: The following steps are involved: Step 1: Data collection and formatting; Real-time collection of highway dynamic data and the length of each road section on the highway; mathematical modeling based on the topological structure of each road section and the highway dynamic data to obtain the road network environment data matrix, road network road condition response data matrix and price control data matrix; Step 2: Machine learning modeling and training; A multi-layer perceptron-based machine learning algorithm is used to extract correlation features and mathematical regression features between the road network environment data matrix, the road network traffic response data matrix, and the price control data matrix to obtain a traffic flow prediction model. Step 3: Multi-objective prediction based on machine learning; By carrying out optimization calculations based on the traffic prediction model, a multi-objective optimization result representing the dynamic toll collection scheme of the expressway is obtained through parallel prediction calculations and objective functions, and a dynamic toll collection scheme of the expressway is recommended based on the multi-objective optimization result. Step 4: Verify and execute the prediction results; Sending the generated multi-objective optimization results representing the dynamic toll collection scheme for the expressway to the administrator for scheme recommendation and obtaining the dynamic toll collection scheme for the expressway selected by the administrator; implementing dynamic toll collection adjustments based on the dynamic toll collection scheme for the expressway selected by the administrator; Step 5: Feedback adjustment of machine learning parameters; Verify the consistency of the multi-objective optimization results, and make feedback corrections to the operating parameters in the flow prediction model based on the verification results to optimize the prediction results of the flow prediction model.

Images